Back to skill

Security audit

Website Capture - AllScreenshots

Security checks for vulnerabilities and agentic risk

Overview

This skill transparently captures website screenshots through a third-party cloud API, with privacy caution needed for sensitive URLs and stored screenshot links.

Before installing, treat Allscreenshots as an external processor: do not submit internal, authenticated, regulated, or secret-bearing pages unless you are authorized and comfortable with the vendor's storage and retention controls. Keep ALLSCREENSHOTS_API_KEY private, and prefer binary output over URL output for sensitive captures when possible.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (7)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill directs users to send target URLs and screenshot requests to a third-party cloud API but does not disclose that requested pages, metadata, and resulting captures are transmitted off-platform for processing. This creates a real privacy and data-handling risk because users may submit internal, sensitive, or authenticated URLs without understanding that the content is being shared with an external service.

External Transmission

Medium
Category
Data Exfiltration
Content
## API Base

Endpoint: `https://api.allscreenshots.com/v1/screenshots`
Auth header: `Bearer $ALLSCREENSHOTS_API_KEY`

## Operations
Confidence
88% confidence
Finding
The documented API base explicitly establishes that all screenshot operations depend on a remote external endpoint, confirming that data will leave the local environment. While this is inherent to the product, it becomes a vulnerability in documentation when there is no accompanying warning about privacy, retention, or transmission of potentially sensitive targets.

External Transmission

Medium
Category
Data Exfiltration
Content
### Desktop screenshot (default)

```bash
curl -s -X POST \
  -H "Authorization: Bearer $ALLSCREENSHOTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"TARGET_URL","fullPage":true,"viewport":{"width":1280,"height":800},"blockAds":true,"blockCookieBanners":true,"stealth":true,"responseType":"url"}' \
Confidence
94% confidence
Finding
This command transmits user-supplied URLs and page-rendering instructions to an external API, which is expected for the skill’s functionality but still constitutes a genuine external data transfer. In this context, the danger is not code execution but that sensitive targets or rendered content may leave the local environment and be processed by a third party, especially with stealth mode intended for bot-protected sites.

External Transmission

Medium
Category
Data Exfiltration
Content
-H "Authorization: Bearer $ALLSCREENSHOTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"TARGET_URL","fullPage":true,"viewport":{"width":1280,"height":800},"blockAds":true,"blockCookieBanners":true,"stealth":true,"responseType":"url"}' \
  "https://api.allscreenshots.com/v1/screenshots" | jq
```

### Mobile screenshot
Confidence
87% confidence
Finding
This example concretely shows transmission of screenshot jobs to the third-party endpoint, including the user-provided URL and capture settings. The skill context makes this somewhat expected, but also more sensitive because screenshots can contain full rendered contents of private pages, not just a simple URL fetch.

External Transmission

Medium
Category
Data Exfiltration
Content
### Mobile screenshot

```bash
curl -s -X POST \
  -H "Authorization: Bearer $ALLSCREENSHOTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"TARGET_URL","fullPage":true,"viewport":{"width":375,"height":812},"deviceScaleFactor":3,"blockAds":true,"blockCookieBanners":true,"stealth":true,"responseType":"url"}' \
Confidence
94% confidence
Finding
The mobile screenshot example also sends target URLs and rendering parameters to an external cloud provider, creating the same privacy and confidentiality exposure as the desktop flow. Because the service renders remote content on behalf of the user, internal or sensitive pages could be disclosed outside the user’s control.

External Transmission

Medium
Category
Data Exfiltration
Content
-H "Authorization: Bearer $ALLSCREENSHOTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"TARGET_URL","fullPage":true,"viewport":{"width":375,"height":812},"deviceScaleFactor":3,"blockAds":true,"blockCookieBanners":true,"stealth":true,"responseType":"url"}' \
  "https://api.allscreenshots.com/v1/screenshots" | jq
```

### Dark mode
Confidence
87% confidence
Finding
The mobile example similarly confirms outbound transmission to the vendor API and therefore the possibility of disclosing sensitive target information or rendered page content. Because the skill advertises stealth mode and bot-protected site support, users may be tempted to use it on more sensitive targets, which increases the practical risk.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
The skill recommends using a CDN-backed storage URL and returning that link to the user, but it does not warn that screenshots may be stored remotely and retrievable via a generated URL. If captures include sensitive page contents, anyone with the link or access to the storage location may be able to view the data, increasing confidentiality and retention risks.

Static analysis

No suspicious patterns detected.